Unanswered: instance hangs - amount of ipcs resources very high
we have DB2 V10.5 FP7 on Linux SLES 11 SP4
quiet regulary two of the five instances are no more available, they hang an no command can be run, all hang
db2 connect to <dbname>
db2 list application
db2 force application all
the only way to be able to work with the database is to kill all db2 processes using ipclean and db2_kill
Before doing that I checked the amount of Memory Segments used by the instance. It is really huge, above 200 whereas the other instances only use max. 100 Memory Segments. This morning there were even more than 1200 segments (ipcs -a | grep <instance_name>
In the db2diag.log I can't find any helpful Information
"ipcs -m" returns memory segments, not "ipcs -a".
What's the size of RAM in the server and what's the output of "ipcs -l"?
If you have an ability to open PMR with IBM, run the following as an instance owner during the hang situation:
db2fodc -hang full
The command collects the information needed to investigate the root cause of the problem.